Unlock instant, AI-driven research and patent intelligence for your innovation.

Accelerate deep neural network in an FPGA

a neural network and acceleration technology, applied in machine learning, multi-programming arrangements, instruments, etc., can solve the problems of not being able to achieve the performance of hardware-centric rtl implementation, and requiring considerable programming effor

Active Publication Date: 2020-05-19
BEIJING PIANRUOJINGHONG TECH CO LTD
View PDF8 Cites 1 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Although some FPGA manufacturers have provided high level synthesis tools that facilitate developers' programming of FPGAs using software-centric programming languages, such as C / C++, Matlab®, and OpenCL®, considerable programming effort remains and the performance of the provided synthesis tools is typically considered not as good as the hardware-centric RTL implementation.

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Accelerate deep neural network in an FPGA
  • Accelerate deep neural network in an FPGA
  • Accelerate deep neural network in an FPGA

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0022]Embodiments of the present invention will now be described in detail with reference to the drawings. However, the following embodiments do not restrict the invention claimed in the claims. Moreover, all features and combinations of features described in various embodiments are not mandatory within the scope and spirit of the present invention. Like numbers are assigned to like components throughout the description of the embodiments of the present invention.

[0023]FIG. 1 illustrates an environment for accelerating a deep neural network (DNN) in an FPGA according to an embodiment of the present invention. As depicted, the environment includes a DNN training platform 10, a DNN conversion platform 20, and a DNN recognition platform 30 including an FPGA 30a.

[0024]In some embodiments, the DNN training platform 10 may be implemented through a graphic processor unit (GPU)-based accelerator (not shown). In some embodiments, the DNN training platform 10 and the DNN conversion platform ...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

A method, system and computer program product for accelerating a deep neural network (DNN) in a field-programmable gate array (FPGA) are disclosed. The method includes receiving a DNN net file and weights, converting the received DNN net file to one or more source files, generating an executable FPGA bit file using the one or more source files, and downloading the executable FPGA bit file from the DNN conversion platform to the FPGA. Converting of the received DNN net file and the weights to the one or more source files can further include analyzing the DNN net file to identify a plurality of neural layers, decomposing one or more neural layers of the plurality of neural layers to one or more operation blocks, instantiating the one or more source files, based on the one or more operation blocks.

Description

FIELD[0001]The present invention relates generally to a deep neural network (DNN), and more particularly to accelerating a DNN in a field-programmable gate array (FPGA).STATEMENT REGARDING PRIOR DISCLOSURES BY THE INVENTOR OR A JOINT INVENTOR[0002]It should be appreciated that non-patent literatures entitled: “Using AccDNN to FPGA-accelerate neural networks without programming”, OpenPOWER™ developer challenge, July 7 (www.youtube.com) and “SuperVessel: cognitive computing platform on the openstack based openpower cloud”, OpenPower™ Summit 2016, Apr. 5-8 were disclosed by the inventor or a joint inventor. Also, these disclosures were made one year or less before the effective filing data of the claimed invention.BACKGROUND[0003]Deep learning has led to state-of-the-art improvements in the accuracy of many artificial intelligence tasks, such as large-category image classification and recognition, speech recognition and nature language processing. The architecture can involve complex a...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
Patent Type & Authority Patents(United States)
IPC IPC(8): G06F9/46G06N3/10G06N3/063
CPCG06F9/46G06N3/105G06N3/063G06N20/00
Inventor LIN, YONGHUATANG, JIANBINWANG, JUNSONG
Owner BEIJING PIANRUOJINGHONG TECH CO LTD